We have developed genetically engineered herpes simplex virus-1 (HSV-1) vectors that can selectively and efficiently infect and kill brain tumor cells in situ without harming surrounding brain cells and without causing systemic disease. Having taken one such vector into human clinical trial, we set forth testable hypotheses aimed at further understanding and improving this method of brain tumor therapy. In order to increase the efficacy of HSV oncolysis in brain tumor therapy, we hypothesize that: a.) HSV oncolytic therapy can be improved by using a HSV vector in conjunction with commonly used chemotherapeutic agents for brain tumors; b.) The efficacy of herpes vectors for brain tumor therapy can be improved through the use of a HSV backbone that replicates better in glioma cells while retaining the necessary safety features for clinical trials. In order to better understand and improve the delivery of HSV vectors for brain tumor therapy, we hypothesize that: a.) Some of the efficacy following intravascular or intratumoral HSV tumor therapy may be due to selective injury of tumor vasculature versus normal vasculature; b.) The timing of co-treatment with antiangiogenesis agents may either inhibit or augment the selective injury to tumor vasculature by oncolytic HSV vectors; c.) Prior anti-HSV immunity could alter the efficacy of intravascular delivery of oncolytic HSV but can be modulated with immunosuppressive agents. In order to further improve the anti-tumor immunity induced by HSV tumor therapy, we hypothesize that: a.) Defective HSV vectors expressing immune-modulatory genes will increase the survival of animals harboring intracranial tumors, b.) A recombinant virus can be constructed from an appropriate parent virus that will express a cytokine without down-regulating MHC-I thus enhancing the anti-tumor immune response.